FinOps Foundation makes the cloud bill everyone's business

Eight people. Eight hours. In early 2020, Pearson put its cloud FinOps team through a Foundation workshop. All eight earned certification. The interesting part came afterward: Ashley Hromatko, then Pearson’s senior cloud FinOps manager, said the team used the discussions to build its roadmap, map milestones to shared principles and borrow automation ideas from other practitioners. A training session had become a way to organize work.

Eight people, eight hours

Pearson’s experience, published when the Linux Foundation announced its intent to host FinOps Foundation in June 2020, captures the organization’s appeal. The difficulty was bigger than reading an invoice. Teams needed a sequence of decisions, a vocabulary and people who had tried the work before. The Foundation gave practitioners a place to compare approaches.

That need had surfaced in Cloudability’s quarterly customer advisory meetings. Practitioners wanted discussions that went beyond vendor tooling. In February 2019, Cloudability helped establish the Foundation, underwriting initial legal and administrative support. Early participants included Spotify, Nationwide and Atlassian. The people buying cloud services had something useful to teach one another.

A profession needs a shared vocabulary

Founding president J.R. Storment came from Cloudability. Mike Fuller, now the Foundation’s CTO, brought practitioner experience from Atlassian. Their backgrounds help explain the emphasis on collaboration: cloud costs emerge from architecture, purchasing and everyday engineering choices. A finance team can report the total without controlling all the decisions that produced it.

The Foundation organizes that problem through its FinOps Framework. Its three phases are Inform, Optimize and Operate: understand usage and cost, evaluate improvements, then implement changes and repeat. Teams can mature individual capabilities gradually. Nobody needs to perfect an entire company before improving one workload.

Its audience includes engineers, finance analysts, procurement specialists and product leaders. Some buy training; organizations can become members; practitioners contribute knowledge. Today the Foundation reports a community exceeding 120,000 people. That measures reach, not a payroll or a list of paying customers. The useful product is the accumulated experience those people can exchange.

Vendor neutrality requires housekeeping. The community vets access by job role, and sales or quota-carrying positions are restricted to Supporter status. Public documentation remains available. These boundaries protect a particular kind of conversation: the engineer explaining what went wrong should not have to sit through a sales pitch first.

The spreadsheet translation tax

There is also a less sociable obstacle. Providers describe billing data differently. Before comparing costs, practitioners may have to reconcile column names, charging conventions and commitment treatment. The translation consumes effort that could have gone into deciding whether a workload earns its keep.

FOCUS, the FinOps Open Cost and Usage Specification, gives billing-data generators common requirements. AWS, Microsoft Azure and Google Cloud are among the providers offering exports. The Foundation supports this work alongside the Framework. One addresses how people cooperate; the other addresses whether their datasets can cooperate.

FOCUS 1.4, ratified in June 2026, adds invoice reconciliation and richer commitment information. That addresses a concrete gap: consumption records do not necessarily contain everything accounts payable needs to match the physical invoice. A shared identifier can be more useful than another chart. It lets two departments check the same transaction.

The organization occupies an unusual market position. Commercial cost platforms sell analysis and automation; cloud providers supply their own billing tools. FinOps Foundation supplies shared practices, learning and standards that can sit underneath those products. Its nonprofit model includes organizational participation, paid education and events. The Framework is publicly accessible, and contributing to FOCUS has no fee, although agreements apply. Certification and implementation still consume money and staff time.

The word that no longer fit

In February 2026, the Foundation changed its mission from advancing people managing cloud value to advancing people managing technology value. Its community had already crossed the boundary. AI, software subscriptions, licensing and data centers were appearing inside the same remit. Keeping “cloud” would have described only part of the work.

The State of FinOps 2026 report found 98% of respondents managing AI spend, compared with 31% two years earlier. The survey represented more than $83 billion in annual cloud spending. These are findings from a practitioner community, not a census of every company. Even with that distinction, the change explains the Foundation’s broader language.

“Advancing the people who manage the value of technology.”The Foundation’s updated mission, 2026

October guidance on AI budget overruns makes the expansion tangible. Model charges sit beside vector databases, orchestration, observability and testing. A budget built only around token prices can overlook the machinery supporting the model. FinOps offers a reason to bring those decisions into one planning conversation before the spending arrives.

Start smaller than the invoice

A reader can copy the approach without creating a department. Pick one workload. Identify who owns it. Agree on a useful business unit, such as a completed transaction, and examine its technology cost. Decide which trade-off matters, implement one change and check the result. This is a practical application of the Framework, rather than a promise of automatic savings.

For billing-data work, the Foundation’s adoption guidance recommends planning, a pilot and education before a wider rollout. Its FOCUS sandbox and free introductory course offer places to begin. Check the version your provider actually exports: publication of a newer specification does not mean every supplier has implemented it.

The remaining constraint is authority. Better reporting cannot make an engineering team prioritize a change or give an analyst permission to alter production. Poor allocation data and absent cooperation can stall the work. Pearson’s example is useful precisely because the workshop continued into a roadmap. The certificate marked a beginning. The next decision was the part that mattered.